SYSTEM ONE MCP · HOSTED CODE MODE
Give your agent Jev.
Let it write the decision flow.
A direct tool answers one bounded question. sysone_code lets your agent ask several Jev questions with loops and conditions, then return the relevant results in one MCP call.
Free early access. Verify your email, add one MCP URL and approve the connection. No payment card or waitlist. Daily limits apply.
A DEPENDENT CHECK INSIDE A LARGER TASK
Use the first answer to choose the next question.
A coding agent searches your project and finds twelve possible causes of a failure. It can ask Jev which excerpts answer the question, check the strongest candidates, and stop when the evidence remains unclear. The agent keeps the original files and decides what to do next.
Code Mode keeps that sequence inside one bounded run. Intermediate results can stay out of the agent's context. See the code and the current limits.
Keep direct tools for one decision.
System One exposes sysone_check, sysone_find, sysone_select, sysone_decide and saved recipes. Each accepts evidence and defined answers. Jev returns choices, scores or probabilities. Use the direct tool when one answer is enough.
Measure the whole task.
Code Mode removes client round trips from a dependent plan but adds a code run. In our local two-step fixture it was slightly slower than two direct calls on the same machine. That fixture used no Jev inference. We have not measured a complete-agent quality, speed or cost gain from Code Mode. See what to compare.
Keep control of the connection.
Every inner call uses the connection's scopes and daily allowance. The code cannot make application actions or direct network requests. Built-in patterns help an agent start with a bounded procedure, and each run has a metadata-only step trace. Connect through OAuth, inspect usage in your workspace, and pause or revoke access when needed.
Hosted Code Mode has a separate daily execution limit. The local application currently offers direct tools with your own provider key.
Create your free workspace →HOSTED MCP
One code tool.
Direct decisions when you need them.
Your agent sees only tools granted to its connection. Code Mode can compose those tools, up to the run's call limit.
Compose a plan
sysone_code runs a bounded JavaScript function with granted jev.* methods.
Check or find
sysone_check tests a claim. sysone_find ranks supplied excerpts.
Choose or run
sysone_select compares options. sysone_run uses one of 45 saved recipes when published to the connection.
The local stdio bridge may list different tool names and does not currently offer Code Mode. Check the tools returned by your connection.
PICK A FIRST TASK
One useful decision.
Your agent does the rest.
Choose a recipe and your agent. Each setup prompt checks for an existing plugin or MCP connection before adding another.
Choose the next tool
Find the code responsible for a form submitting twice.
How it helps
Jev selects from the available tool descriptions. The agent supplies the arguments and runs the tool.
Give a repeated routing choice to a small model and keep the main agent on the repair.
Search found several files. Which ones explain this bug?
How it helps
Jev rates the retrieved passages against the task. The original passages stay available for review.
Reduce irrelevant context before the next reasoning step.
Check a claim before repeating it
Does this tool output support the claim that the release is live?
How it helps
Jev checks the statement against supplied evidence and separates contradiction from missing information.
Catch an unsupported conclusion before it becomes an answer or another action.
Match a palette to a brief
Choose a calm palette for a coastal reading app.
How it helps
Jev compares your described palettes with the brief. The agent applies the chosen tokens and checks contrast in code.
Reuse a selection recipe for design choices with a clear brief and a defined set of options.
Find the item a person describes
Find a compact desk lamp with warm light and no subscription.
How it helps
Jev compares the supplied item descriptions and returns a candidate or a reason to review the choice.
Match natural language to an existing catalog without generating an item that does not exist.
Choose a useful diagnostic check
The service is running, but requests still fail. What should we inspect next?
How it helps
Jev chooses among the checks the agent can actually run, using the symptoms already collected.
Make the next investigation step explicit. Compare it with a fixed diagnostic sequence.